- Email: [email protected]
Surface EMG modifications after eccentric exercise
J. Elecrromyogr.
Kinesiol.
0 1997 ELSEVIER
Elsevier
Vol. 7, No. 3, pp. 193-202, 1997 Science Ltd. All rights reserved Printed in Great Britain 105lM411/97 $17.00 + 0.00
PII: SlOSO-6411(96)00034-X
Surface EMG Modifications
After Eccentric Exercise
Francesco Felici’T2, Lorenzo Colace’ and Paola Sbriccoli’ ‘Institute
of Human Physiology, Faculty of Medicine, lJniversit2 di Rorna ‘La Sapienza’, Rome, Italy; ‘I.R.C.C.S. Rehabilitation Hospital, S. Lucia, Rome, Italy
Summary: The possibility that the surface electromyographic signal (sEMG) from exercised muscle would show significant changes to demonstrate muscle damage after eccentric contraction (EC) was tested in this study. The experiment lasted five consecutive days. On the first day, six sedentary adult subjects performed two rounds of 35 ECs with the biceps brachii of the non-dominant arm, the other arm being used as control. Individual muscle soreness was assessed on a subjective scale. The analysis of sEMG was performed on the signal recorded during isometric contractions at 80% and 50% of maximal voluntary contraction (MVC), choosing root mean square (RMS) and median frequency (MDF) as synthetic sEMG parameters. MVC was also recorded, and plasma levels of creatine kinase were determined in four subjects. The most important findings which resulted from this study were: (a) spectral parameters are less sensitive to error introduced by electrode repositioning than time domain parameters, and are more sensitive to EC-induced sEMG changes than RMS; (b) a significant shift of MDF power spectra towards low frequencies at 80% and 50% MVC (20% and 5% of decay, respectively) was evident as early as 1 h after EC on the exercised arm; and (c) MDF follows the evolution of muscle damage. We concluded from these results that MDF is suitable for the early and non-invasive detection of sEMG changes induced by EC. In addition, we found further evidence that the observed modifications result from a selective or prevalent damage of type 2B muscle fibres. 0 1997 Elsevier Science Ltd Key Words: Muscle damage-Biceps brachii-Surface Median frequency-Spectral compression.
electromyography-
J. Electromyogr. Kinesiol., Vol. 7, 193-202, September
INTRODUCTION
More severe damage is promoted in sedentary subjects by stretching the muscle while it is actively shortening (eccentric contraction; EC). Typical symptoms of this damage are represented by the delayed onset of muscle soreness, which usually appears 36-48 h after the trauma, by a consistent decrease of the maximal voluntary force (MVC) attainable in the isometric condition and by the accompanying swelling and oedema of muscle and surrounding tissues. Muscle damage is associated with an increase of creatine kinase (CK) plasma levels which peak 2448 h after the trauma, and
It is well known that strenuous physical exercise can produce more or less marked muscle damage6J4. The extent and severity of the damage are dependent upon the kind of muscle action performed and the previous level of muscular training of the subjecP9. Received 17 September 1996. Correspondence and reprint requests to F. Felici, Istituto di Fisiologia Umana, Facolti di Medicina e Chirurgia, Universita degli Studi di Roma ‘La Sapienza’, Piazzale Aldo Moro 5, 00185 Roma, Italy.
193
194
F. FELICI
which are more pronounced following EC than after concentric or isometric exercises6T7. Muscle biopsies provided direct evidence of morphological and histochemical changes induced within the muscle by EC*. Friden and Liebe?, Newhamr8 and Friden et a1.9, using light and electron microscopy, showed that fast glycolitic muscle fibres (type 2B) were more affected than slow oxidative (type 1) and fast oxidative (type 2A) muscle fibres. Newham18 reported the presence of small focal areas of fibre damage immediately after EC, which become more extensive during the following 2-3 days. The damage started with Z-lines disruption, then continued with myofibrillar disruption and the spreading of Z-lines material throughout the sarcomere. Besides all these changes, which have already been shown to be associated with eccentric exercise, it seems that structural damage of muscle fibres promoted by EC should cause, in turn, a concomitant modification of the electric activity generated by the muscle. The analysis of the surface electromyographic signal (sEMG) could conveniently be used to detect which changes in the electric behaviour of a muscle, if any, are associated with EC. The sEMG during voluntary isometric contractions is commonly synthetically described by means of two parameters*7.21: the root mean square (RMS; time domain) and the median frequency (MDF; frequency domain). It has been proposed1,4,23 that the RMS and MDF provide information relative to the number and location of the active motor units, the recruitment of motor units, the shape of the motor unit action potentials, the mean firing rate of the individual motor units, and the extent of superposition of action potentials from concurrently active motor units. Although it is very difficult to isolate the individual contribution of any of the influential factors listed from the sEMG analysis, it is conceivable that when one or more of the signal generators are damaged, this would be reflected within the synthetic parameter considered. It is worthy of note that there have only been two reports so far describing the EC exercise-induced modifications of the surface myoelectric signal **I*. These results are also, to some extent, controversial. Berry et al.* compared eccentric with concentric contractions. They found the increase of the sEMG RMS to be greater after EC than after concentric exercise, but did not observe any significant differences in the MPF values before and after EC, nor between EC and concentric contractions. They limited their observations to 48 h after the EC, whereas it is well known that in this
ET AL.
condition subjective and objective signs of muscle damage are continuing long after this time”. On the other hand, Kroon and Naeije’* showed that the rate of MPF decline along with contraction time was steeper after EC rather than after concentric and isometric exercises. These parameters recovered within 10 days. In this study the MPF,, i.e. the MPF at the start of each contraction epoch, remained unchanged before and after EC-induced muscle damage. They also observed that the increase of sEMG RMS was greater after eccentric exercise than after the other two contraction modalities. We designed the present study to investigate further: (a) if there is any significant and discernible modification of sEMG after EC; (b) in the case of sEMG changes, if they are caused by EC or unwanted muscle fatigue induced by the experimental protocol adopted; (c) if these changes are eventually related to indices of muscle damage (onset and evolution of muscle pain, force decrease, plasma CK increase) so that it would be possible to establish a causal relationship between these changes and muscle damage. In particular, we explored the sEMG focusing on RMS and MDF, aiming at the evaluation of their usefulness in terms of their consistency during different days and their sensitivity to EC-induced myoelectric modifications. We defined consistency as the reproducibility of results obtained on different days, whereas sensitivity is the early detection of myoelectric signal modifications that may be attributed to EC. In order to make our results comparable with those obtained by other authors, we adopted the experimental protocol proposed by Clarkson3, and we studied the biceps brachii of both sides, exercising only one side with EC and using the other as control. METHODS Six adult subjects (three males and three females, aged 29.2f5 yr; body weight 64k12.2 kg; height 168.3f9.3 cm (mean + SD) who did not practise any kind of systematic muscular training took part in the experiment. All subjects provided an informed written consent. At the time of the experiment, all subjects were in good physical condition with no signs or symptoms of neuromuscular disease. They were not under pharmacological treatments and did not diet. All the subjects were requested not to take any extra exercise involving the arm muscles during the whole period of the experiment. All subjects were right handed and the biceps brachii of the
EMG AFTER MUSCLE DAMAGE
non-dominant arm (the left one in all instances) underwent eccentric exercise, whereas the biceps brachii of the dominant arm was used as control. A modified version of the subjective pain rate scale proposed by Berry et al2 was given to each subject. Obviously, we did not include any comparison between exercised arms. Every day, before starting the exercise phase of the experiment, subjects were asked to report their subjective sensation of muscle pain on the scale which ranged from normal to very, very sore in 10 steps. Four subjects (three males and one female) consented to the determination of their CK plasma levels. Blood samples were drawn by venipuncture of the anterocubital vein before the beginning of each set of measurements. Samples were collected in a B-D ethylenediamine tetraacetic acid (EDTA) K3 vacutainer containing 0.07 ml of 15% EDTA. Plasma was obtained after 10 min centrifugation; CK plasma levels were assessed using Sigma test kit 47-10 (Sigma Chemical Co., Italy) at a temperature of 30°C. Total test time was 5 min; expected normal values using this particular test should range from 15 to 110 U 1-r for men and from 15 to 90 U 1-l for women. Subjects performed isometric and eccentric contractions, and sEMG and force signals were recorded during isometric and eccentric contractions. The analysis of sEMG was only performed on isometric recordings to ensure that distortion introduced by relative movements between electrodes and muscle would be minimal. Isometric and eccentric exercises were performed with the mechanical apparatus illustrated in Figure 1. The subject arm position was accurately recorded by drawing an arm silhouette on the inclined plane to ensure a correct subject repositioning during the 5 days of the experiment. Particular care was taken to mark the exact position of the wrist strap to provide an accurate assessment of force over contractions, i.e. to avoid errors due to lever arm changing. Wrist strap positioning was determined by having the superior edge of the strap coincide with a horizontal line joining the two stiloideous processes of the radius and ulna. This line was traced with dermographic ink on the skin. During the isometric contraction the apparatus was set to obtain a fixed elbow angle of 90”. During the eccentric contraction, the chain, C2, was removed and the subject was pulling with all his/her might while the lever, L, was moved in the opposite direction by an operator, performing the stretching of the muscle. In this case the elbow angle ranged
195
FIG. 1. Sketch of the mechanical device used for isometric and eccentric contractions. During EC the chain, C2, was removed. Cl-C2: chains; T: force transducer; L: lever.
from 30” to 170”. The chain, Cl, was equipped with a force transducer (Kistler, type 93llA; sensitivity -3.93 PC/N, linearity <0.3%) connected with a charge amplifier Kistler 5001 (cut-off frequency set at 180 kHz). The experimental protocol consisted of the following phases (see also Table 1): Phase 1: Left and right biceps. MVC was measured for 5 s. Three measurements were taken at intervals of 3 min. The highest value was taken as representing the 100% MVC. Phase 2: Left and right biceps. The subject performed three isometric contractions at 80% MVC lasting 12 s and spaced by 5 min of rest. Then, after 15 min of rest, three contractions were performed at 50% MVC of equal duration and spaced as above. Phase 3: Left biceps only. The subject performed the eccentric exercise. The eccentric exercise protocol reproduced that proposed by Clarkson and consisted of two rounds of 35 contractions (3 s of contraction and 12 s of passive recovery to the starting position). The two rounds were separated by a 5-min interval. Phase 4. Left and right biceps. One hour after the end of EC the same set of measurements as in Phases 1 and 2 were repeated. Phases 1 and 2 and phases 2 and 3 were spaced by 15 min. It should be noted that phases 3 and 4 were performed only on Monday. Therefore, during
196
F. FELICI TABLE
1.
ET AL.
Summary of the experimental protocol Phase
Right biceps Elbow angle measurement MVC measurement (3x5 s) 3 min rest between
each
test
1 Left biceps Elbow angle measurement MVC measurement (3x5s)
rest between
each
test
Phase 15 min rest 80% MVC measurement 15 min rest 50% MVC measurement
(3x12
s) 5 min rest between
each
test
(3x12
s) 5 min rest between
each
test
2 15 min rest 80% MVC measurement 15 min rest 50% MVC measurement
3min
(3x12
s) 5 min rest between
each
test
(3x12
s) 5 min rest between
each
test
-
Phase
3 Left biceps First bout of 35 eccentric contractions each lasting 15 s (3 s stretching and 12 s passive return to starting position) 5 min rest Second bout of 35 eccentric contractions each lasting 15 s (3 s stretching and 12 s passive return to starting position)
Same
as in phases
1 and 2; performed
1 h after
CK was sampled before phase 1 and before exercise was performed only on the first day.
phase phase
Phase 4 3 on the first day 3 on the
the following 4 days, the experimental protocol was reduced to phases 1 and 2. At the end of the experimental week, a total of 18 recordings at 80% and 18 at 50% MVC were obtained for each arm and for each subject. sEMG surface electrodes (3 M) were Ag/AgCl 10 mm diameter on self-adhesive support. The electrodes were positioned on the middle portion of the muscle belly (short head) parallel to the longitudinal axis of the muscle fibres, with an inter-electrode distance (centre-to-centre) of 20 mm. The contours of the self-adhesive plastic discs were marked on the skin using dermographic ink. In the following 4 days, these skin markers were used to place the new electrodes as close as possible to the original position. The sEMG signal was handled by an HP8811A amplifier with the pass band set between 1.5 Hz and 1 kHz. The acquisition system consisted of a multichannel PC board with a 1Zbit successive approximation A/D converter. The sampling frequency was 2048 Hz. Calibration of the A/D converter and data acquisition were software controlled. Quantitative analysis was only performed on sEMGs recorded during 80% and 50% MVC isometric contractions. Each parameter was calculated over subsequent epochs of 1 s of sEMG signal. The analysis was performed starting from the third second of signal to skip transient phenomena. Because each test was repeated three times at any given level of force, for any second of sEMG signal,
first
day,
only and
before
phase
1 on the following
days.
Eccentric
each parameter value resulted from the mean of three values. The RMS is analytically defined as:
sEMG,,,.& t) = where x(t) is the sEMG signal and T is the acquisition time. The evaluation of the spectral parameters was performed by means of a standard fast Fourier transform (FFT) over 2048 samples. The MDF, that frequency value which bisects the spectrum into two regions of equal power, is defined as:
MDF.TPmdf
= j. Pcf)df = ; jPind/
0
0
fmed
where P(f) is the power spectrum density. Statistical
Analysis
The following description refers to all RMS and MDF data obtained from contractions at 80% and 50% MVC on both arms. Statistical analyses of all considered parameters were performed. In accordance with Merletti et a1.15, linear regression analysis (least squares method) of the time course of RMS
EMG AFTER MUSCLE DAMAGE and MDF was performed on data collected during every experiment. We tested the null hypothesis of linearity of each regression by means of an F-test*O. The null hypothesis was retained at a level of confidence PKO.05 with a tabulated value of F~8k20;o;05)=2.45.An example of this test is reported in Figure 2 for the right biceps MDF and BMS data obtained on Monday before EC (Man*) and Friday (Fri). The null hypothesis of an identity between linear regressions before and after EC was tested both
197
intra- and inter-individually following the procedure described by SachszO. To reduce possible contamination of the sEMG signal by fatigue phenomena, the axis intercept at time 0 @MS0 and MDF,) was assumed to be indicative of the initial state of muscle activation. Because we discarded the first 2 s of every contraction, this is an extrapolation which may slightly underestimate the real intercepP5. To make comparisons of the two sides of a subject and between different subjects, the slopes and intercepts were
FRI
MO/V* 120
MDF = 96.~1.9.nme
MDF = 83.65 -1.71~l-d F&l= 0394.45 I% O.G5
Ft&=o.25c2.45
p
;;ilOO ru & 80 zz 60
40I
40
0
2
4
6
8
10
12
Time [s]
FRI
MON* 600 -,
1
Rys=24&94+1.72llme
-.s. 6 ‘5
500400-
$
8== 3clo -.- ..-_____,__ .. ... .__ ...&..).A .._......-.--....m ______ ____,_ :r-..;...:..-’ .. .... .w-t..p..,
0 0
I 2
I 4
I 6 Time [s]
0 W
First test Second Test
A
Third test
I 8
I
I
10
12
0
RM~ = 440.05 Ftet= n.s.
2
4
- 13.6 Time
6
8
10
12
Time [s]
FIG. 2. Examples of plot of MDF (upper panels) and RMS (lower panels) vs. time during 80% WC isometric contractions on the right (control) biceps brachii. In each graph, different symbols refer to the three contractions performed. Regression lines, 95% confidence interval, regression equations and f-test results are reported. NS: not significant. All contractions are from the same subject. Mon.: first control test; Fri: sixth control test.
198
F. FELJCI ET AL..
normalized with respect to their corresponding values obtained before EC and a r-test for paired data adopted.
RESULTS One hour after eccentric exercise, a 45% reduction of mean MVC was observed with respect to initial values. No significant recovery was present throughout the duration of experiment (see Figure 3A); furthermore, after EC, the mean MVC values did not differ from each other over the remaining days of the experiment. On the control arm, the MVC did not show significant departures from control values. In no instance did any subject report a clear feeling of pain in the exercised muscle, but all subjects uniformly reported a sensation of moderate soreness which peaked between 24 and 72 h after the exercise (scores between 3 and 5 on the proposed scale2) and that remained constant throughout the experimental week. CK levels always started to move significantly from the baseline 24-48 h after exercise and continued to rise until Friday (Fri)
A
I
I
I
I
Tue
Wed
Thu
Fri
-70L--d/ Mona
Mon
Time [hs]
Time [days]
4000
B y
3000
1
7
oI/
I Mon’
Mon
Time [hs]
Tua
I
I
I
Wed
Thu
Fri
Time [days]
FIG. 3. Time course of left biceps at 100% MVC during experimental week. Mean and standard error values from six subjects are reported. Mon samples were obtained after EC.
the the 1h
(2241 U l-‘f35.8 U l-l, mean rt SE; see Figure 3B); it was impossible to show a statistically significant relationship between the onset of subjective sensation of muscle impairment - or its peak - and CK modifications or between loss of force and CK level modifications (compare data presented in Figure 3A and B). As already stated, we discarded the first 2 s of each contraction in order to avoid transient phenomena; RMS and MDF were thus computed over the remaining 10 s on epochs of 1 s. Figure 2 (upper panel) reports a representative example of the time course of MDF values for the right arm along with the correspondent regression analysis. MDF was computed on spectra obtained by the three daily contractions performed at 80% MVC (first control test; Man*) and 4 days after (sixth control test; Fri). The MDF linearity test reported in this Figure was confirmed in more than 96% of the curves obtained from the six subjects, i.e. on 140 out of 144 curves (24 regressions per subject considering two force levels and two sides). The RMS analysis did not provide results of similar quality, more than 70% of the data being not statistically linear (Figure 2, lower panel). As a consequence, no further statistical tests were applied to RMS data. Nevertheless, it is worthy of note that on the exercised arm the RhIS slope increased sharply in all subjects 1 h after the EC at 80% and 50% MVC but such an effect vanished 1 or 2 days after the EC. For each subject, the identity tests performed on MDF curves produced the following results: (a) on the right biceps at 80% and 50% MVC, all the individual curves obtained after the Man* reference test were not statistically different from the reference curve itself (PCO.05); (B) on the left biceps at 80% and 50% MVC, the intercepts of the curves obtained 1 h after the EC were statistically different from the reference values (PcO.05) with Man* intercept greater than Mon. This result was confirmed during the following days. The statistical analysis on the rate of MDF decay showed that linear regressions before and after EC were not different (P
199
EMG AFTER MUSCLE DAMAGE EC. In Figure 4, all the MDF data obtained at 80% MVC from the six subjects are presented for the right and left sides, whereas in Figure 5 data at 50% MVC are reported. Data before EC were compared with their correspondent post-EC data. MDF, of pre- and post-exercise curves on the left side showed a significant difference at the pre-set level of confidence both at 80% and 50% MVC. Also in this case, the absence of any significant difference between the slopes of the curves was confirmed. Bight side data are organized in the same way as for the left side. The excellent repeatability of these data in any condition and the absolute identity of the compared pairs of data seem clear. 100 90
Before EC: 8297 (n=l00)
ARer EC: 69.51 - O.WTime
(n=em) 80
T
I 401 0
2
4
6
8
10
12
Time [s] 100
1
0 0
90
Before EC: MDF = 84.94 - 1 .SSlIme (n-100)
B&m (tl=leq
EC: 83.82 - O.WTime : 01.16 - O.Sl.Time
idler EC: MDF = 64.29 - Z .56linw 80
80 i? I
9
5;
-1.04lime
L
70
70
P
5
60
60
50
50 40
I 0
1 2
I 4
I 6
I a
I 10
40
I 12
0
100
k
4
6
a
10
12
Time [s]
Time [s]
is I
2
90
0
Before EC: MDF = 70.6 - 1.75.Time (ll=lrn)
00
0
After EC: MDF = 77.01-
1 .AB.Time
70
5
FIG. 5. Upper panel: left biceps MDF mean data and standard errors (six subjects) vs. time before EC (circles) and after EC (squares) at 50% MVC. After EC data were grouped from 1 h after to Friday. Regression lines and 95% confidence interval are shown along with their corresponding equations. The number n indicates the samples used to compute the regression equations. Lower panel: same as above for the right biceps. In this case before EC data were arbitrarily separated from after EC data for the sake of comparison with left biceps experiments.
60
0
2
4
6
a
10
12
Time [s]
FIG. 4. Upper panel: left biceps MDF mean data and standard errors (six subjects) vs. time before EC (circles) and after EC (squares) at 80% MVC. After EC data were grouped from 1 h after to Friday. Regression lines and 95% confidence interval are shown along with their corresponding equations. The number n indicates the samples used to compute the regression equations. Lower panel: same as above for the right biceps. In this case before EC data were arbitrarily separated from after EC data for the sake of comparison with left biceps experiments.
In Figure 6, the daily MDFo values were computed as the mean of the intercepts of all subjects and normalized with respect to the corresponding pre-exercise value. MDF, was significantly affected by EC both at 80% and 50%. Although the same overall results can be observed at 80% and 50% MVC force levels, it appears that the decrease of MDF, at the lowest force level is more delayed and less evident. Besides, at 80% MVC the mean intercept values were not statistically different (t-test, P> 0.05) from each other through the duration of the experiment (from Mon to Fri).
200
F. FELICI
130
0
1
q
Left arm Right arm
I 701, Mot-P
Mon
Time
130
T
115
zi g -1
100
i5 3
B z
I Tue
I Wed
, Thu
I Fri
Time [days]
[hs]
1 1
--p--
---
a
1
f
---p----L
P
85
P
1 70&/I Man’
8 Mon
Tue
I Wed
I Thu
I Fri
Time [days]
Time [hs]
FIG. 6. Mean (six subjects) normalized MDF intercept values (MDF,) obtained on different days at 80% MVC (upper panel) and 50% MVC (lower panel). MDFo values were normalized with respect to their before EC data.
DISCUSSION This study presents convincing indications that in repeated measurements (a) the EC causes significant changes of sEMG; (b) these changes are not an artefact caused by the experimental protocol adopted. When all the external factors are carefully kept constant, MDF is a stable sEMG parameter and is less subject than RMS (see Figure 2) to errors caused by, for example, electrode repositioning, among many possible causes of signal distortion. Because one of the main goals of this work was to identify reproducible sEMG parameters to characterize signal modification, we thus preferred MDF. Another important criterion in making this decision was the possibility of expressing parameter evolution in a simple form. Although RMS is generally
ET AL.
accepted as a good indicator of some properties of the sEMG signa14, its representation is often more complicated than a linear function of time and, in fact, in the present work the statistical test of RMS linear dependence on the time of isometric contraction was very often rejected. In addition, in cases such as that presented in Figure 2, the general aspect of the RMS time course changed significantly from Man* to Fri. Merletti et a1.16 reported greater oscillations of RMS initial values than those of MDF during electrically elicited contractions, either before or after electrode repositioning. It was also noted that inter-day variability was greater than intra-day variability. Our results confirm those of Merletti et a1.16 and lead us to the conclusion that the added noise introduced by electrode repositioning influences time domain much more than frequency domain parameters. We also obtained a noticeable repeatability of MDF, on the control arm which was within a range of +2% of the reference value (see Figure 6). This result is considerably better than those reported by othersi6, and is well within the error range associated with the estimate of frequency parameters of sEMG power spectra”. The EC performed following the protocol of Clarkson et a1.3 induced effective muscle damage as confirmed by CK modifications measured in our subjects. When MVC measurements were repeated on the control arm, we did not observe any significant modification of MVC values from the reference contraction. This result, along with the fact that during the following 4 days the MVC remained at a constant fraction of the original value without any clear sign of recovery, indicates that the isometric exercises included in our experimental protocol did not induce a significant amount of fatigue. On the other hand, this points to the conclusion that the MVC reduction on the left biceps was caused by EC. The lack of correlation between the onset of CK modifications and MVC decrease is not new3.5g7*‘8,19,24, nor is that between CK and the onset of muscle soreness. Both are explained by the time needed before the damage may affect the sarcolemmatic structure of muscle fibres. An original finding presented in this paper is that no causal relationship seems to exist between a rise in CK levels and sEMG modifications, other than that they are co-existent but not coincident in time. If the structural damage promoted by EC is causing the change in sEMG (as stated in the Introduction) force and sEMG values should follow the CK levels. At 80% MVC force level, after an immediate
EMG AFTER MUSCLE
decrease of both frequency and force, there is a tendency for the MDFo and MVC to be constant. A somewhat different behaviour is observed at 50% MVC. Beginning 48 h after EC, MDF, starts to decrease at a more or less constant rate in a fashion which closely resembles that of CK increase. One must be extremely cautious in interpreting such indirect results as being indicative of a causal relationship between muscle damage and sEMG power spectra changes, especially when it is not possible to observe the damage directly on muscle biopsies. Nevertheless, the possibility cannot be excluded that, for reasons that are yet to be explained, immediately after EC there is a dramatic reduction of force and MDF which is not paralleled by an increase of CK nor by an immediate painful sensation. These results are in contrast with two previous studies*,‘* which, to our knowledge, are the only recent papers in which an attempt was made to characterize EC by means of sEMG analysis. The above discrepancy requires that several factors be discussed. A general aspect that could have produced differences in the MDF, between this and the other reports, could be that different amounts of muscle damage were promoted. It has been proposed in animal models’4 that the extent of muscle damage following EC is highly correlated with the strain imposed on the muscle during active lengthening and, to a lesser extent, with the absolute value of force developed during the EC. Although we cannot present quantitative data on muscle strain, we may at least state that damage has been promoted (CK increase). It is also easy to calculate that during EC we imposed an angular velocity of 47” s-i, which is well above the value presented by the classical work of Singh and Karpovich22. The exercise protocol we adopted here is therefore certainly capable of producing a significant active strain on the exercised muscle. A comparison with the report of Kroon and Naeije’* reveals that: (a) MVC recovered at more than 80% of reference on the fourth day; and (b) the muscle load during EC was set to 40% of the subject’s MVC, whereas in our experiment subjects exercised at between 120% and 70% MVC. Kroon and Naeije’*, while adopting the least squares method, obtained their MDF, at 40% MVC. We have clearly shown (see Figure 4 and Figure 5) that a great difference exists between results obtained at 80% and 50% MVC on MDF, and it is even greater when data are normalized with respect to reference values (Figure 6). The lack of observed modifications in the normalized initial MDF at 40% MVC
DAMAGE
201
is, thus, not surprising considering which factors are thought to influence MDF. It has been proposed1*4v23 that MDF contains information relative to the recruitment of motor units, the shape of the motor unit action potentials, the mean firing rate of the individual motor units and the extent of superposition of action potentials from concurrently active motor units. There is a widespread belief that, in muscles such as the biceps brachii, the recruitment of new motor units continues until almost 80% MVC is reached’3s23. In single motor unit studies5 there is evidence that any newly recruited motor unit needs time to reach a steady state firing rate. On the basis of the so called ‘common drive of motor units’5 it is conceivable that when working at 40% MVC, even if high threshold fast twitch motor units were active, they would not be firing at their maximum rate but would eventually be driven at the low firing rates typical of the small motor neurons. Berry et a1.2 did not present any MVC data. A more solid explanation for the discrepancy with the results of these authors may be found in the different methodological approaches used. There is a great difference in the way we obtained MDFo and the MPF values presented by the above cited authors. In the present paper, we computed MDF on time epochs of 1 s and then, using a least squares approximation method, we extrapolated the MDF,; whereas Berry et a1.2 presented a mean MPF value resulting from the average of 10 MPF subsequent samples. The latter procedure should have exerted a smoothing effect on the computed parameter; in addition, signal analysis might have been corrupted by the fact that sEMG was recorded during active concentric or eccentric muscle action and not during isometric contraction. This can be assimilated to a random electrode repositioning procedure that might have masked MPF modifications induced by exercise. In conclusion, we can state that: (a) when an accurate experimental procedure is adopted, it is possible to obtain reproducible results from sEMG analysis, particularly with respect to the frequency domain; (b) synthetic parameters obtained by the frequency domain analysis, such as the MDF, may be conveniently adopted to extract information from sEMG after EC analogous to that observable during muscle fatigue experiments; (c) muscle damage after EC is non-invasively detected by means of sEMG analysis. From this point of view, typical manifestations of such damage are represented by an early
202
F. FELICI ET AL.
decrease of the very initial frequency content of sEMG, which accurately follows the evolution of damage. Our results, confirming those obtained by other researchers, add further evidence to the hypothesis that it is type 2B muscle fibres which are predominantly affected by EC. If this is confirmed, for example by sEMG recordings and muscle biopsy specimens collected in the same experiment, attempts may be made to apply this technique to the quantitative functional evaluation of muscle training procedures which specifically include EC. Finally, one may postulate that, as causative factors of MDF compression, action potentials from motor neurons innervating fast twitch fibres are suppressed and/or reduced, and/or there is an increased synchronization of motor unit action potentials. In this case, it is necessary to clarify two points: what information is sent to the central nervous system to activate such an inhibitory network and, second, where does it originate. The authors wish to thank Prof. Bernard0 Pensa from the Department of Biochemistry of University of Rome ‘La Sapienza’ for plasma CK level determination and Prof. Marco Marchetti, Director of the Institute of Human Physiology of the University of Rome ‘La Sapienza’ for his valuable help and advice in the preparation of the manuscript. This work was supported by the Faculty of Medicine of Rome University ‘La Sapienza’ and by IRCCs, St Lucia, Rome, Italy.
6 7
8. 9. 10. 11. 12. 13.
14. 15.
Acknowledgements:
16. 17. 18. 19.
REFERENCES 1. Bemardi M, Solomonow M, Sanchez JH, Baratta RV, Nguyen G: Motor unit recruitment strategy of knee antagonist muscles in a step-wise, increasing isometric contraction. Eur J Appl Physiol 70:493-501, 1995. 2. Berry CB, Moritani T, Tolson H: Electrical activity and soreness in muscle after exercise. Am J Phys Med Rehabil 69:6&66, 1990. 3. Clarkson PM, Nosaka K, Braun B: Muscle function after exercise-induced muscle damage and rapid adaptation. Med Sci Sports Exert 24:512-520, 1992. 4. De Luca CJ, Knaflitz M: Surfnce Electromyogruphy: What’s New? CLUT Edizioni Pub., Torino, Italy, 1992. 5. De Luca CJ, Erim Z: Common drive of motor units in
20. 21. 22. 23.
24.
regulation of muscle force. Trends Neurosci 17:299-305, 1994. Ebbeling CB, Clarkson PM: Exercise-induced muscle damage and adaptation. Sports Med 7~207-234, 1989. Evans WJ. Meredith CN. Cannon JG. Dinarello CA. Frontera WR, Hughes VA, Jones BH, Knuttgen HG: Metabolic changes following eccentric exercise in trained and untrained men. J Appl Physiol 61:1864-1868, 1986. Friden J, Lieber RL: Structural and mechanical basis of exercise-induced muscle injury. Med Sci Sports Exert 24:521-530, 1992. Friden J, Seger J, Siostrom M, Ekblom B: Adaptive response in human skeletal muscle subjected to prolonged eccentric trainina. Int J Soorts Med 4:177-183. 1983. Friden”J, Sjostrom M, Ekblom B: Myofibrillar damage following intense eccentric exercise in man. lnt I Sports Med 4:170-176, 1983. Hof L: Errors in frequency parameters of EMG power spectra. IEEE Truns Biomed Engng 38:1077-1088, 1991. Kroon GW, Naeije M: Recovery of the human biceps electromyogram after heavy eccentric, concentric or isometric exercise. Eur J Appl Physiol 63:444-448, 1991. Kukulka CG, Clamann HP: Comparison of the recruitment and discharge properties of motor units in human brachial biceps and adductor pollicis during isometric contractions. Brain Res 219:45-55, 1981. Lieber R, Friden J: Muscle damage is not a function of muscle force but active muscle strain. J Appl Physiol 74:520526, 1993. Merletti R, Knaflitz M, De Luca CJ: Myoelectric manifestations of fatigue in voluntary and electrically elicited contractions. J Azwl Phvsiol 69:1810-1820. 1990. Merletti R,‘Lo Cbnte LR, Sathyan D: Repeatability of electrically evoked myoelectric signals in the human tibialis anterior muscle. J Electromyogr Kinesiol 5:67-80, 1995. Merletti R, Lo Conte LR, Orizio C: Indices of muscle fatigue. J Elecromyogr Kinesiol 1:20-33, 1991. Newham DJ: The consequences of eccentric contractions and their relationship to delayed onset muscle pain. Eur J Appl Physiol 57:353-359, 1988. Nosaka K, Clarkson PM, McGuiggin ME, Byrne JM: Time course of muscle adaptation after high force eccentric exercise. Eur J Appl Physiol 63:7(X76, 1991. Sachs L: Applied Statistics. A Handbook of Techniques, 2nd edn, Springer Verlag, New York, 1984. Sargeant AJ, Kernel1 D: Neuromuscolar fatigue. Proceedings of the Symposium, Amsterdam, 9-11 April 1992. NorthHolland Publ., Amsterdam, 1993. Singh M, Karpovich V: Isotonic and isometric forces of forearm flexors and extensors. J Appl Physiol 21:14351437, 1966. Solomonow M, Baten C, Smit J, Baratta R, Hermens H, D’ Ambrosia R, Shoji H: Electromyogram power spectra frequencies associated with motor unit recruitment strategies. / Appl Physiol 68: 1177-I 185, 1990. Stauber WT: Eccentric action of muscles: physiology, injury, and adaptation. Ex Sport Sci Rev 17:157-185, 1989.
Kinesiol.
0 1997 ELSEVIER
Elsevier
Vol. 7, No. 3, pp. 193-202, 1997 Science Ltd. All rights reserved Printed in Great Britain 105lM411/97 $17.00 + 0.00
PII: SlOSO-6411(96)00034-X
Surface EMG Modifications
After Eccentric Exercise
Francesco Felici’T2, Lorenzo Colace’ and Paola Sbriccoli’ ‘Institute
of Human Physiology, Faculty of Medicine, lJniversit2 di Rorna ‘La Sapienza’, Rome, Italy; ‘I.R.C.C.S. Rehabilitation Hospital, S. Lucia, Rome, Italy
Summary: The possibility that the surface electromyographic signal (sEMG) from exercised muscle would show significant changes to demonstrate muscle damage after eccentric contraction (EC) was tested in this study. The experiment lasted five consecutive days. On the first day, six sedentary adult subjects performed two rounds of 35 ECs with the biceps brachii of the non-dominant arm, the other arm being used as control. Individual muscle soreness was assessed on a subjective scale. The analysis of sEMG was performed on the signal recorded during isometric contractions at 80% and 50% of maximal voluntary contraction (MVC), choosing root mean square (RMS) and median frequency (MDF) as synthetic sEMG parameters. MVC was also recorded, and plasma levels of creatine kinase were determined in four subjects. The most important findings which resulted from this study were: (a) spectral parameters are less sensitive to error introduced by electrode repositioning than time domain parameters, and are more sensitive to EC-induced sEMG changes than RMS; (b) a significant shift of MDF power spectra towards low frequencies at 80% and 50% MVC (20% and 5% of decay, respectively) was evident as early as 1 h after EC on the exercised arm; and (c) MDF follows the evolution of muscle damage. We concluded from these results that MDF is suitable for the early and non-invasive detection of sEMG changes induced by EC. In addition, we found further evidence that the observed modifications result from a selective or prevalent damage of type 2B muscle fibres. 0 1997 Elsevier Science Ltd Key Words: Muscle damage-Biceps brachii-Surface Median frequency-Spectral compression.
electromyography-
J. Electromyogr. Kinesiol., Vol. 7, 193-202, September
INTRODUCTION
More severe damage is promoted in sedentary subjects by stretching the muscle while it is actively shortening (eccentric contraction; EC). Typical symptoms of this damage are represented by the delayed onset of muscle soreness, which usually appears 36-48 h after the trauma, by a consistent decrease of the maximal voluntary force (MVC) attainable in the isometric condition and by the accompanying swelling and oedema of muscle and surrounding tissues. Muscle damage is associated with an increase of creatine kinase (CK) plasma levels which peak 2448 h after the trauma, and
It is well known that strenuous physical exercise can produce more or less marked muscle damage6J4. The extent and severity of the damage are dependent upon the kind of muscle action performed and the previous level of muscular training of the subjecP9. Received 17 September 1996. Correspondence and reprint requests to F. Felici, Istituto di Fisiologia Umana, Facolti di Medicina e Chirurgia, Universita degli Studi di Roma ‘La Sapienza’, Piazzale Aldo Moro 5, 00185 Roma, Italy.
193
194
F. FELICI
which are more pronounced following EC than after concentric or isometric exercises6T7. Muscle biopsies provided direct evidence of morphological and histochemical changes induced within the muscle by EC*. Friden and Liebe?, Newhamr8 and Friden et a1.9, using light and electron microscopy, showed that fast glycolitic muscle fibres (type 2B) were more affected than slow oxidative (type 1) and fast oxidative (type 2A) muscle fibres. Newham18 reported the presence of small focal areas of fibre damage immediately after EC, which become more extensive during the following 2-3 days. The damage started with Z-lines disruption, then continued with myofibrillar disruption and the spreading of Z-lines material throughout the sarcomere. Besides all these changes, which have already been shown to be associated with eccentric exercise, it seems that structural damage of muscle fibres promoted by EC should cause, in turn, a concomitant modification of the electric activity generated by the muscle. The analysis of the surface electromyographic signal (sEMG) could conveniently be used to detect which changes in the electric behaviour of a muscle, if any, are associated with EC. The sEMG during voluntary isometric contractions is commonly synthetically described by means of two parameters*7.21: the root mean square (RMS; time domain) and the median frequency (MDF; frequency domain). It has been proposed1,4,23 that the RMS and MDF provide information relative to the number and location of the active motor units, the recruitment of motor units, the shape of the motor unit action potentials, the mean firing rate of the individual motor units, and the extent of superposition of action potentials from concurrently active motor units. Although it is very difficult to isolate the individual contribution of any of the influential factors listed from the sEMG analysis, it is conceivable that when one or more of the signal generators are damaged, this would be reflected within the synthetic parameter considered. It is worthy of note that there have only been two reports so far describing the EC exercise-induced modifications of the surface myoelectric signal **I*. These results are also, to some extent, controversial. Berry et al.* compared eccentric with concentric contractions. They found the increase of the sEMG RMS to be greater after EC than after concentric exercise, but did not observe any significant differences in the MPF values before and after EC, nor between EC and concentric contractions. They limited their observations to 48 h after the EC, whereas it is well known that in this
ET AL.
condition subjective and objective signs of muscle damage are continuing long after this time”. On the other hand, Kroon and Naeije’* showed that the rate of MPF decline along with contraction time was steeper after EC rather than after concentric and isometric exercises. These parameters recovered within 10 days. In this study the MPF,, i.e. the MPF at the start of each contraction epoch, remained unchanged before and after EC-induced muscle damage. They also observed that the increase of sEMG RMS was greater after eccentric exercise than after the other two contraction modalities. We designed the present study to investigate further: (a) if there is any significant and discernible modification of sEMG after EC; (b) in the case of sEMG changes, if they are caused by EC or unwanted muscle fatigue induced by the experimental protocol adopted; (c) if these changes are eventually related to indices of muscle damage (onset and evolution of muscle pain, force decrease, plasma CK increase) so that it would be possible to establish a causal relationship between these changes and muscle damage. In particular, we explored the sEMG focusing on RMS and MDF, aiming at the evaluation of their usefulness in terms of their consistency during different days and their sensitivity to EC-induced myoelectric modifications. We defined consistency as the reproducibility of results obtained on different days, whereas sensitivity is the early detection of myoelectric signal modifications that may be attributed to EC. In order to make our results comparable with those obtained by other authors, we adopted the experimental protocol proposed by Clarkson3, and we studied the biceps brachii of both sides, exercising only one side with EC and using the other as control. METHODS Six adult subjects (three males and three females, aged 29.2f5 yr; body weight 64k12.2 kg; height 168.3f9.3 cm (mean + SD) who did not practise any kind of systematic muscular training took part in the experiment. All subjects provided an informed written consent. At the time of the experiment, all subjects were in good physical condition with no signs or symptoms of neuromuscular disease. They were not under pharmacological treatments and did not diet. All the subjects were requested not to take any extra exercise involving the arm muscles during the whole period of the experiment. All subjects were right handed and the biceps brachii of the
EMG AFTER MUSCLE DAMAGE
non-dominant arm (the left one in all instances) underwent eccentric exercise, whereas the biceps brachii of the dominant arm was used as control. A modified version of the subjective pain rate scale proposed by Berry et al2 was given to each subject. Obviously, we did not include any comparison between exercised arms. Every day, before starting the exercise phase of the experiment, subjects were asked to report their subjective sensation of muscle pain on the scale which ranged from normal to very, very sore in 10 steps. Four subjects (three males and one female) consented to the determination of their CK plasma levels. Blood samples were drawn by venipuncture of the anterocubital vein before the beginning of each set of measurements. Samples were collected in a B-D ethylenediamine tetraacetic acid (EDTA) K3 vacutainer containing 0.07 ml of 15% EDTA. Plasma was obtained after 10 min centrifugation; CK plasma levels were assessed using Sigma test kit 47-10 (Sigma Chemical Co., Italy) at a temperature of 30°C. Total test time was 5 min; expected normal values using this particular test should range from 15 to 110 U 1-r for men and from 15 to 90 U 1-l for women. Subjects performed isometric and eccentric contractions, and sEMG and force signals were recorded during isometric and eccentric contractions. The analysis of sEMG was only performed on isometric recordings to ensure that distortion introduced by relative movements between electrodes and muscle would be minimal. Isometric and eccentric exercises were performed with the mechanical apparatus illustrated in Figure 1. The subject arm position was accurately recorded by drawing an arm silhouette on the inclined plane to ensure a correct subject repositioning during the 5 days of the experiment. Particular care was taken to mark the exact position of the wrist strap to provide an accurate assessment of force over contractions, i.e. to avoid errors due to lever arm changing. Wrist strap positioning was determined by having the superior edge of the strap coincide with a horizontal line joining the two stiloideous processes of the radius and ulna. This line was traced with dermographic ink on the skin. During the isometric contraction the apparatus was set to obtain a fixed elbow angle of 90”. During the eccentric contraction, the chain, C2, was removed and the subject was pulling with all his/her might while the lever, L, was moved in the opposite direction by an operator, performing the stretching of the muscle. In this case the elbow angle ranged
195
FIG. 1. Sketch of the mechanical device used for isometric and eccentric contractions. During EC the chain, C2, was removed. Cl-C2: chains; T: force transducer; L: lever.
from 30” to 170”. The chain, Cl, was equipped with a force transducer (Kistler, type 93llA; sensitivity -3.93 PC/N, linearity <0.3%) connected with a charge amplifier Kistler 5001 (cut-off frequency set at 180 kHz). The experimental protocol consisted of the following phases (see also Table 1): Phase 1: Left and right biceps. MVC was measured for 5 s. Three measurements were taken at intervals of 3 min. The highest value was taken as representing the 100% MVC. Phase 2: Left and right biceps. The subject performed three isometric contractions at 80% MVC lasting 12 s and spaced by 5 min of rest. Then, after 15 min of rest, three contractions were performed at 50% MVC of equal duration and spaced as above. Phase 3: Left biceps only. The subject performed the eccentric exercise. The eccentric exercise protocol reproduced that proposed by Clarkson and consisted of two rounds of 35 contractions (3 s of contraction and 12 s of passive recovery to the starting position). The two rounds were separated by a 5-min interval. Phase 4. Left and right biceps. One hour after the end of EC the same set of measurements as in Phases 1 and 2 were repeated. Phases 1 and 2 and phases 2 and 3 were spaced by 15 min. It should be noted that phases 3 and 4 were performed only on Monday. Therefore, during
196
F. FELICI TABLE
1.
ET AL.
Summary of the experimental protocol Phase
Right biceps Elbow angle measurement MVC measurement (3x5 s) 3 min rest between
each
test
1 Left biceps Elbow angle measurement MVC measurement (3x5s)
rest between
each
test
Phase 15 min rest 80% MVC measurement 15 min rest 50% MVC measurement
(3x12
s) 5 min rest between
each
test
(3x12
s) 5 min rest between
each
test
2 15 min rest 80% MVC measurement 15 min rest 50% MVC measurement
3min
(3x12
s) 5 min rest between
each
test
(3x12
s) 5 min rest between
each
test
-
Phase
3 Left biceps First bout of 35 eccentric contractions each lasting 15 s (3 s stretching and 12 s passive return to starting position) 5 min rest Second bout of 35 eccentric contractions each lasting 15 s (3 s stretching and 12 s passive return to starting position)
Same
as in phases
1 and 2; performed
1 h after
CK was sampled before phase 1 and before exercise was performed only on the first day.
phase phase
Phase 4 3 on the first day 3 on the
the following 4 days, the experimental protocol was reduced to phases 1 and 2. At the end of the experimental week, a total of 18 recordings at 80% and 18 at 50% MVC were obtained for each arm and for each subject. sEMG surface electrodes (3 M) were Ag/AgCl 10 mm diameter on self-adhesive support. The electrodes were positioned on the middle portion of the muscle belly (short head) parallel to the longitudinal axis of the muscle fibres, with an inter-electrode distance (centre-to-centre) of 20 mm. The contours of the self-adhesive plastic discs were marked on the skin using dermographic ink. In the following 4 days, these skin markers were used to place the new electrodes as close as possible to the original position. The sEMG signal was handled by an HP8811A amplifier with the pass band set between 1.5 Hz and 1 kHz. The acquisition system consisted of a multichannel PC board with a 1Zbit successive approximation A/D converter. The sampling frequency was 2048 Hz. Calibration of the A/D converter and data acquisition were software controlled. Quantitative analysis was only performed on sEMGs recorded during 80% and 50% MVC isometric contractions. Each parameter was calculated over subsequent epochs of 1 s of sEMG signal. The analysis was performed starting from the third second of signal to skip transient phenomena. Because each test was repeated three times at any given level of force, for any second of sEMG signal,
first
day,
only and
before
phase
1 on the following
days.
Eccentric
each parameter value resulted from the mean of three values. The RMS is analytically defined as:
sEMG,,,.& t) = where x(t) is the sEMG signal and T is the acquisition time. The evaluation of the spectral parameters was performed by means of a standard fast Fourier transform (FFT) over 2048 samples. The MDF, that frequency value which bisects the spectrum into two regions of equal power, is defined as:
MDF.TPmdf
= j. Pcf)df = ; jPind/
0
0
fmed
where P(f) is the power spectrum density. Statistical
Analysis
The following description refers to all RMS and MDF data obtained from contractions at 80% and 50% MVC on both arms. Statistical analyses of all considered parameters were performed. In accordance with Merletti et a1.15, linear regression analysis (least squares method) of the time course of RMS
EMG AFTER MUSCLE DAMAGE and MDF was performed on data collected during every experiment. We tested the null hypothesis of linearity of each regression by means of an F-test*O. The null hypothesis was retained at a level of confidence PKO.05 with a tabulated value of F~8k20;o;05)=2.45.An example of this test is reported in Figure 2 for the right biceps MDF and BMS data obtained on Monday before EC (Man*) and Friday (Fri). The null hypothesis of an identity between linear regressions before and after EC was tested both
197
intra- and inter-individually following the procedure described by SachszO. To reduce possible contamination of the sEMG signal by fatigue phenomena, the axis intercept at time 0 @MS0 and MDF,) was assumed to be indicative of the initial state of muscle activation. Because we discarded the first 2 s of every contraction, this is an extrapolation which may slightly underestimate the real intercepP5. To make comparisons of the two sides of a subject and between different subjects, the slopes and intercepts were
FRI
MO/V* 120
MDF = 96.~1.9.nme
MDF = 83.65 -1.71~l-d F&l= 0394.45 I% O.G5
Ft&=o.25c2.45
p
;;ilOO ru & 80 zz 60
40I
40
0
2
4
6
8
10
12
Time [s]
FRI
MON* 600 -,
1
Rys=24&94+1.72llme
-.s. 6 ‘5
500400-
$
8== 3clo -.- ..-_____,__ .. ... .__ ...&..).A .._......-.--....m ______ ____,_ :r-..;...:..-’ .. .... .w-t..p..,
0 0
I 2
I 4
I 6 Time [s]
0 W
First test Second Test
A
Third test
I 8
I
I
10
12
0
RM~ = 440.05 Ftet= n.s.
2
4
- 13.6 Time
6
8
10
12
Time [s]
FIG. 2. Examples of plot of MDF (upper panels) and RMS (lower panels) vs. time during 80% WC isometric contractions on the right (control) biceps brachii. In each graph, different symbols refer to the three contractions performed. Regression lines, 95% confidence interval, regression equations and f-test results are reported. NS: not significant. All contractions are from the same subject. Mon.: first control test; Fri: sixth control test.
198
F. FELJCI ET AL..
normalized with respect to their corresponding values obtained before EC and a r-test for paired data adopted.
RESULTS One hour after eccentric exercise, a 45% reduction of mean MVC was observed with respect to initial values. No significant recovery was present throughout the duration of experiment (see Figure 3A); furthermore, after EC, the mean MVC values did not differ from each other over the remaining days of the experiment. On the control arm, the MVC did not show significant departures from control values. In no instance did any subject report a clear feeling of pain in the exercised muscle, but all subjects uniformly reported a sensation of moderate soreness which peaked between 24 and 72 h after the exercise (scores between 3 and 5 on the proposed scale2) and that remained constant throughout the experimental week. CK levels always started to move significantly from the baseline 24-48 h after exercise and continued to rise until Friday (Fri)
A
I
I
I
I
Tue
Wed
Thu
Fri
-70L--d/ Mona
Mon
Time [hs]
Time [days]
4000
B y
3000
1
7
oI/
I Mon’
Mon
Time [hs]
Tua
I
I
I
Wed
Thu
Fri
Time [days]
FIG. 3. Time course of left biceps at 100% MVC during experimental week. Mean and standard error values from six subjects are reported. Mon samples were obtained after EC.
the the 1h
(2241 U l-‘f35.8 U l-l, mean rt SE; see Figure 3B); it was impossible to show a statistically significant relationship between the onset of subjective sensation of muscle impairment - or its peak - and CK modifications or between loss of force and CK level modifications (compare data presented in Figure 3A and B). As already stated, we discarded the first 2 s of each contraction in order to avoid transient phenomena; RMS and MDF were thus computed over the remaining 10 s on epochs of 1 s. Figure 2 (upper panel) reports a representative example of the time course of MDF values for the right arm along with the correspondent regression analysis. MDF was computed on spectra obtained by the three daily contractions performed at 80% MVC (first control test; Man*) and 4 days after (sixth control test; Fri). The MDF linearity test reported in this Figure was confirmed in more than 96% of the curves obtained from the six subjects, i.e. on 140 out of 144 curves (24 regressions per subject considering two force levels and two sides). The RMS analysis did not provide results of similar quality, more than 70% of the data being not statistically linear (Figure 2, lower panel). As a consequence, no further statistical tests were applied to RMS data. Nevertheless, it is worthy of note that on the exercised arm the RhIS slope increased sharply in all subjects 1 h after the EC at 80% and 50% MVC but such an effect vanished 1 or 2 days after the EC. For each subject, the identity tests performed on MDF curves produced the following results: (a) on the right biceps at 80% and 50% MVC, all the individual curves obtained after the Man* reference test were not statistically different from the reference curve itself (PCO.05); (B) on the left biceps at 80% and 50% MVC, the intercepts of the curves obtained 1 h after the EC were statistically different from the reference values (PcO.05) with Man* intercept greater than Mon. This result was confirmed during the following days. The statistical analysis on the rate of MDF decay showed that linear regressions before and after EC were not different (P
199
EMG AFTER MUSCLE DAMAGE EC. In Figure 4, all the MDF data obtained at 80% MVC from the six subjects are presented for the right and left sides, whereas in Figure 5 data at 50% MVC are reported. Data before EC were compared with their correspondent post-EC data. MDF, of pre- and post-exercise curves on the left side showed a significant difference at the pre-set level of confidence both at 80% and 50% MVC. Also in this case, the absence of any significant difference between the slopes of the curves was confirmed. Bight side data are organized in the same way as for the left side. The excellent repeatability of these data in any condition and the absolute identity of the compared pairs of data seem clear. 100 90
Before EC: 8297 (n=l00)
ARer EC: 69.51 - O.WTime
(n=em) 80
T
I 401 0
2
4
6
8
10
12
Time [s] 100
1
0 0
90
Before EC: MDF = 84.94 - 1 .SSlIme (n-100)
B&m (tl=leq
EC: 83.82 - O.WTime : 01.16 - O.Sl.Time
idler EC: MDF = 64.29 - Z .56linw 80
80 i? I
9
5;
-1.04lime
L
70
70
P
5
60
60
50
50 40
I 0
1 2
I 4
I 6
I a
I 10
40
I 12
0
100
k
4
6
a
10
12
Time [s]
Time [s]
is I
2
90
0
Before EC: MDF = 70.6 - 1.75.Time (ll=lrn)
00
0
After EC: MDF = 77.01-
1 .AB.Time
70
5
FIG. 5. Upper panel: left biceps MDF mean data and standard errors (six subjects) vs. time before EC (circles) and after EC (squares) at 50% MVC. After EC data were grouped from 1 h after to Friday. Regression lines and 95% confidence interval are shown along with their corresponding equations. The number n indicates the samples used to compute the regression equations. Lower panel: same as above for the right biceps. In this case before EC data were arbitrarily separated from after EC data for the sake of comparison with left biceps experiments.
60
0
2
4
6
a
10
12
Time [s]
FIG. 4. Upper panel: left biceps MDF mean data and standard errors (six subjects) vs. time before EC (circles) and after EC (squares) at 80% MVC. After EC data were grouped from 1 h after to Friday. Regression lines and 95% confidence interval are shown along with their corresponding equations. The number n indicates the samples used to compute the regression equations. Lower panel: same as above for the right biceps. In this case before EC data were arbitrarily separated from after EC data for the sake of comparison with left biceps experiments.
In Figure 6, the daily MDFo values were computed as the mean of the intercepts of all subjects and normalized with respect to the corresponding pre-exercise value. MDF, was significantly affected by EC both at 80% and 50%. Although the same overall results can be observed at 80% and 50% MVC force levels, it appears that the decrease of MDF, at the lowest force level is more delayed and less evident. Besides, at 80% MVC the mean intercept values were not statistically different (t-test, P> 0.05) from each other through the duration of the experiment (from Mon to Fri).
200
F. FELICI
130
0
1
q
Left arm Right arm
I 701, Mot-P
Mon
Time
130
T
115
zi g -1
100
i5 3
B z
I Tue
I Wed
, Thu
I Fri
Time [days]
[hs]
1 1
--p--
---
a
1
f
---p----L
P
85
P
1 70&/I Man’
8 Mon
Tue
I Wed
I Thu
I Fri
Time [days]
Time [hs]
FIG. 6. Mean (six subjects) normalized MDF intercept values (MDF,) obtained on different days at 80% MVC (upper panel) and 50% MVC (lower panel). MDFo values were normalized with respect to their before EC data.
DISCUSSION This study presents convincing indications that in repeated measurements (a) the EC causes significant changes of sEMG; (b) these changes are not an artefact caused by the experimental protocol adopted. When all the external factors are carefully kept constant, MDF is a stable sEMG parameter and is less subject than RMS (see Figure 2) to errors caused by, for example, electrode repositioning, among many possible causes of signal distortion. Because one of the main goals of this work was to identify reproducible sEMG parameters to characterize signal modification, we thus preferred MDF. Another important criterion in making this decision was the possibility of expressing parameter evolution in a simple form. Although RMS is generally
ET AL.
accepted as a good indicator of some properties of the sEMG signa14, its representation is often more complicated than a linear function of time and, in fact, in the present work the statistical test of RMS linear dependence on the time of isometric contraction was very often rejected. In addition, in cases such as that presented in Figure 2, the general aspect of the RMS time course changed significantly from Man* to Fri. Merletti et a1.16 reported greater oscillations of RMS initial values than those of MDF during electrically elicited contractions, either before or after electrode repositioning. It was also noted that inter-day variability was greater than intra-day variability. Our results confirm those of Merletti et a1.16 and lead us to the conclusion that the added noise introduced by electrode repositioning influences time domain much more than frequency domain parameters. We also obtained a noticeable repeatability of MDF, on the control arm which was within a range of +2% of the reference value (see Figure 6). This result is considerably better than those reported by othersi6, and is well within the error range associated with the estimate of frequency parameters of sEMG power spectra”. The EC performed following the protocol of Clarkson et a1.3 induced effective muscle damage as confirmed by CK modifications measured in our subjects. When MVC measurements were repeated on the control arm, we did not observe any significant modification of MVC values from the reference contraction. This result, along with the fact that during the following 4 days the MVC remained at a constant fraction of the original value without any clear sign of recovery, indicates that the isometric exercises included in our experimental protocol did not induce a significant amount of fatigue. On the other hand, this points to the conclusion that the MVC reduction on the left biceps was caused by EC. The lack of correlation between the onset of CK modifications and MVC decrease is not new3.5g7*‘8,19,24, nor is that between CK and the onset of muscle soreness. Both are explained by the time needed before the damage may affect the sarcolemmatic structure of muscle fibres. An original finding presented in this paper is that no causal relationship seems to exist between a rise in CK levels and sEMG modifications, other than that they are co-existent but not coincident in time. If the structural damage promoted by EC is causing the change in sEMG (as stated in the Introduction) force and sEMG values should follow the CK levels. At 80% MVC force level, after an immediate
EMG AFTER MUSCLE
decrease of both frequency and force, there is a tendency for the MDFo and MVC to be constant. A somewhat different behaviour is observed at 50% MVC. Beginning 48 h after EC, MDF, starts to decrease at a more or less constant rate in a fashion which closely resembles that of CK increase. One must be extremely cautious in interpreting such indirect results as being indicative of a causal relationship between muscle damage and sEMG power spectra changes, especially when it is not possible to observe the damage directly on muscle biopsies. Nevertheless, the possibility cannot be excluded that, for reasons that are yet to be explained, immediately after EC there is a dramatic reduction of force and MDF which is not paralleled by an increase of CK nor by an immediate painful sensation. These results are in contrast with two previous studies*,‘* which, to our knowledge, are the only recent papers in which an attempt was made to characterize EC by means of sEMG analysis. The above discrepancy requires that several factors be discussed. A general aspect that could have produced differences in the MDF, between this and the other reports, could be that different amounts of muscle damage were promoted. It has been proposed in animal models’4 that the extent of muscle damage following EC is highly correlated with the strain imposed on the muscle during active lengthening and, to a lesser extent, with the absolute value of force developed during the EC. Although we cannot present quantitative data on muscle strain, we may at least state that damage has been promoted (CK increase). It is also easy to calculate that during EC we imposed an angular velocity of 47” s-i, which is well above the value presented by the classical work of Singh and Karpovich22. The exercise protocol we adopted here is therefore certainly capable of producing a significant active strain on the exercised muscle. A comparison with the report of Kroon and Naeije’* reveals that: (a) MVC recovered at more than 80% of reference on the fourth day; and (b) the muscle load during EC was set to 40% of the subject’s MVC, whereas in our experiment subjects exercised at between 120% and 70% MVC. Kroon and Naeije’*, while adopting the least squares method, obtained their MDF, at 40% MVC. We have clearly shown (see Figure 4 and Figure 5) that a great difference exists between results obtained at 80% and 50% MVC on MDF, and it is even greater when data are normalized with respect to reference values (Figure 6). The lack of observed modifications in the normalized initial MDF at 40% MVC
DAMAGE
201
is, thus, not surprising considering which factors are thought to influence MDF. It has been proposed1*4v23 that MDF contains information relative to the recruitment of motor units, the shape of the motor unit action potentials, the mean firing rate of the individual motor units and the extent of superposition of action potentials from concurrently active motor units. There is a widespread belief that, in muscles such as the biceps brachii, the recruitment of new motor units continues until almost 80% MVC is reached’3s23. In single motor unit studies5 there is evidence that any newly recruited motor unit needs time to reach a steady state firing rate. On the basis of the so called ‘common drive of motor units’5 it is conceivable that when working at 40% MVC, even if high threshold fast twitch motor units were active, they would not be firing at their maximum rate but would eventually be driven at the low firing rates typical of the small motor neurons. Berry et a1.2 did not present any MVC data. A more solid explanation for the discrepancy with the results of these authors may be found in the different methodological approaches used. There is a great difference in the way we obtained MDFo and the MPF values presented by the above cited authors. In the present paper, we computed MDF on time epochs of 1 s and then, using a least squares approximation method, we extrapolated the MDF,; whereas Berry et a1.2 presented a mean MPF value resulting from the average of 10 MPF subsequent samples. The latter procedure should have exerted a smoothing effect on the computed parameter; in addition, signal analysis might have been corrupted by the fact that sEMG was recorded during active concentric or eccentric muscle action and not during isometric contraction. This can be assimilated to a random electrode repositioning procedure that might have masked MPF modifications induced by exercise. In conclusion, we can state that: (a) when an accurate experimental procedure is adopted, it is possible to obtain reproducible results from sEMG analysis, particularly with respect to the frequency domain; (b) synthetic parameters obtained by the frequency domain analysis, such as the MDF, may be conveniently adopted to extract information from sEMG after EC analogous to that observable during muscle fatigue experiments; (c) muscle damage after EC is non-invasively detected by means of sEMG analysis. From this point of view, typical manifestations of such damage are represented by an early
202
F. FELICI ET AL.
decrease of the very initial frequency content of sEMG, which accurately follows the evolution of damage. Our results, confirming those obtained by other researchers, add further evidence to the hypothesis that it is type 2B muscle fibres which are predominantly affected by EC. If this is confirmed, for example by sEMG recordings and muscle biopsy specimens collected in the same experiment, attempts may be made to apply this technique to the quantitative functional evaluation of muscle training procedures which specifically include EC. Finally, one may postulate that, as causative factors of MDF compression, action potentials from motor neurons innervating fast twitch fibres are suppressed and/or reduced, and/or there is an increased synchronization of motor unit action potentials. In this case, it is necessary to clarify two points: what information is sent to the central nervous system to activate such an inhibitory network and, second, where does it originate. The authors wish to thank Prof. Bernard0 Pensa from the Department of Biochemistry of University of Rome ‘La Sapienza’ for plasma CK level determination and Prof. Marco Marchetti, Director of the Institute of Human Physiology of the University of Rome ‘La Sapienza’ for his valuable help and advice in the preparation of the manuscript. This work was supported by the Faculty of Medicine of Rome University ‘La Sapienza’ and by IRCCs, St Lucia, Rome, Italy.
6 7
8. 9. 10. 11. 12. 13.
14. 15.
Acknowledgements:
16. 17. 18. 19.
REFERENCES 1. Bemardi M, Solomonow M, Sanchez JH, Baratta RV, Nguyen G: Motor unit recruitment strategy of knee antagonist muscles in a step-wise, increasing isometric contraction. Eur J Appl Physiol 70:493-501, 1995. 2. Berry CB, Moritani T, Tolson H: Electrical activity and soreness in muscle after exercise. Am J Phys Med Rehabil 69:6&66, 1990. 3. Clarkson PM, Nosaka K, Braun B: Muscle function after exercise-induced muscle damage and rapid adaptation. Med Sci Sports Exert 24:512-520, 1992. 4. De Luca CJ, Knaflitz M: Surfnce Electromyogruphy: What’s New? CLUT Edizioni Pub., Torino, Italy, 1992. 5. De Luca CJ, Erim Z: Common drive of motor units in
20. 21. 22. 23.
24.
regulation of muscle force. Trends Neurosci 17:299-305, 1994. Ebbeling CB, Clarkson PM: Exercise-induced muscle damage and adaptation. Sports Med 7~207-234, 1989. Evans WJ. Meredith CN. Cannon JG. Dinarello CA. Frontera WR, Hughes VA, Jones BH, Knuttgen HG: Metabolic changes following eccentric exercise in trained and untrained men. J Appl Physiol 61:1864-1868, 1986. Friden J, Lieber RL: Structural and mechanical basis of exercise-induced muscle injury. Med Sci Sports Exert 24:521-530, 1992. Friden J, Seger J, Siostrom M, Ekblom B: Adaptive response in human skeletal muscle subjected to prolonged eccentric trainina. Int J Soorts Med 4:177-183. 1983. Friden”J, Sjostrom M, Ekblom B: Myofibrillar damage following intense eccentric exercise in man. lnt I Sports Med 4:170-176, 1983. Hof L: Errors in frequency parameters of EMG power spectra. IEEE Truns Biomed Engng 38:1077-1088, 1991. Kroon GW, Naeije M: Recovery of the human biceps electromyogram after heavy eccentric, concentric or isometric exercise. Eur J Appl Physiol 63:444-448, 1991. Kukulka CG, Clamann HP: Comparison of the recruitment and discharge properties of motor units in human brachial biceps and adductor pollicis during isometric contractions. Brain Res 219:45-55, 1981. Lieber R, Friden J: Muscle damage is not a function of muscle force but active muscle strain. J Appl Physiol 74:520526, 1993. Merletti R, Knaflitz M, De Luca CJ: Myoelectric manifestations of fatigue in voluntary and electrically elicited contractions. J Azwl Phvsiol 69:1810-1820. 1990. Merletti R,‘Lo Cbnte LR, Sathyan D: Repeatability of electrically evoked myoelectric signals in the human tibialis anterior muscle. J Electromyogr Kinesiol 5:67-80, 1995. Merletti R, Lo Conte LR, Orizio C: Indices of muscle fatigue. J Elecromyogr Kinesiol 1:20-33, 1991. Newham DJ: The consequences of eccentric contractions and their relationship to delayed onset muscle pain. Eur J Appl Physiol 57:353-359, 1988. Nosaka K, Clarkson PM, McGuiggin ME, Byrne JM: Time course of muscle adaptation after high force eccentric exercise. Eur J Appl Physiol 63:7(X76, 1991. Sachs L: Applied Statistics. A Handbook of Techniques, 2nd edn, Springer Verlag, New York, 1984. Sargeant AJ, Kernel1 D: Neuromuscolar fatigue. Proceedings of the Symposium, Amsterdam, 9-11 April 1992. NorthHolland Publ., Amsterdam, 1993. Singh M, Karpovich V: Isotonic and isometric forces of forearm flexors and extensors. J Appl Physiol 21:14351437, 1966. Solomonow M, Baten C, Smit J, Baratta R, Hermens H, D’ Ambrosia R, Shoji H: Electromyogram power spectra frequencies associated with motor unit recruitment strategies. / Appl Physiol 68: 1177-I 185, 1990. Stauber WT: Eccentric action of muscles: physiology, injury, and adaptation. Ex Sport Sci Rev 17:157-185, 1989.